Institute for Renewable Energy

Short description of the position:

Eurac Research is looking for a postdoctoral researcher with a background in energy system modelling and/or machine learning to join the Overall Energy System Modelling and e-mobility research group of the Institute for Renewable Energy. The position is part of SURE-RES (Spatially-resolved UnceRtainty analysis for futurE Renewable based Energy Systems), a three-year research project carried out jointly with the Climate Policy Lab of ETH Zurich and co-funded by the Swiss National Science Foundation (SNSF) and the Autonomous Province of Bolzano.

SURE-RES develops multi-fidelity methods that make rigorous uncertainty analysis possible in high-resolution energy system models. Starting from a sector-coupled PyPSA-Eur model of Europe with about 200 nodes and hourly resolution, the project develops and compares two strategies: resolution-reduced models combined with polynomial chaos expansion, and graph neural network (GNN) surrogates that emulate the hourly operational optimisation while preserving full spatial detail. The results are translated into robust, spatially-resolved policy guidance for South Tyrol and Switzerland.

You will be one of the main researchers of the Eurac team in the project, working on the GNN surrogate and resolution-reduced models, on the comparison of the multi-fidelity strategies and on the policy analysis, in close collaboration with the ETH Zurich team. The work combines methodological research at the frontier of energy modelling and artificial intelligence with real-world applications for energy planning

Tasks:

- Contribute, together with ETH Zurich,



to the development and validation of the high-fidelity sector-coupled energy system model of Europe in PyPSA-Eur, with detailed representation of South Tyrol and Switzerland, including data collection, definition of policy assumptions and of the uncertainty parameter space.
- Build resolution-reduced model variants (spatial, temporal and sector-coupling aggregation), train multi-fidelity polynomial chaos expansion surrogates and perform global sensitivity analysis (e.g. Sobol indices).
- Design, train and validate graph neural network surrogates (e.g. GAT, GraphSAGE, GIN) of the hourly operational optimisation and integrate them into a capacity expansion planning framework.
- Compare the multi-fidelity strategies in terms of accuracy, computational efficiency, spatial detail and uncertainty propagation, and produce spatially-resolved uncertainty assessments and policy recommendations for South Tyrol and Switzerland.
- Write peer-reviewed publications and project reports, present results at international conferences and release open-source code and datasets.
- Contribute to the coordination with ETH Zurich (regular online and in-person meetings, reporting), support the supervision of PhD students and interns, and contribute to new research proposals.

Requirements:

- PhD in Energy Engineering, Computer Science, Applied Mathematics,



Physics or a related field (candidates close to completing their PhD may also be considered).
- Proven experience in energy system modelling and optimisation (e.g. PyPSA, Calliope, oemof, TIMES or similar) and/or in machine learning applied to energy or engineering systems.
- Very good programming skills in Python, including scientific computing and data analysis.
- Track record of peer-reviewed publications in relevant journals.
- Very good command of English, written and spoken; knowledge of Italian and/or German is a plus.
- experience with PyPSA/PyPSA-Eur and linear optimisation solvers (e.g. Gurobi, HiGHS);
- experience with deep learning frameworks (e.g. PyTorch, PyTorch Geometric) and graph neural networks;
- knowledge of uncertainty quantification and global sensitivity analysis methods (e.g. polynomial chaos expansion, Sobol indices, Monte Carlo methods);
- experience with high-performance computing, Linux environments and version control (Git);
- knowledge of explainable AI techniques (e.g. SHAP);
- experience in proposal writing, student supervision or interaction with policy makers and stakeholders;
- Italian and/or German.

We offer:

- A fixed-term contract within the SURE-RES project (duration: 1-3 years).
- The opportunity to work on a research topic at the frontier of energy system modelling and artificial intelligence, with direct relevance for real-world energy planning.
- Close collaboration with the Climate Policy Lab of ETH Zurich, including regular in-person meetings in Zurich and Bolzano.
- Working conditions that allow professional growth and know-how acquisition, with career opportunities as a researcher.
- The chance to work with local and international

📌 Postdoctoral Researcher in Energy System Modelling and Machine Learning (Italia)
🏢 Institute for Biomedicine – Eurac Research
📍 Italia